Engines of Growth : Essays in Swedish Economic History

LUND UNIVERSITY PO Box 117

Engines of Growth

Essays in Swedish Economic History

Berger, Thor

2016

Document Version:

Publisher's PDF, also known as Version of record

Link to publication

Citation for published version (APA):

Berger, T. (2016). Engines of Growth: Essays in Swedish Economic History.

Total number of authors:

1

General rights

Unless other specific re-use rights are stated the following general rights apply:

Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.

• Users may download and print one copy of any publication from the public portal for the purpose of private study or research.

• You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portal

Read more about Creative commons licenses: https://creativecommons.org/licenses/ Take down policy

If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Engines of Growth

Essays in Swedish Economic History

Thor Berger

Distribueras av

Media-Tryck, Lunds universitet

Box 117, 221 00 Lund

bookorder@se.lu.se

Thor Berger

Engines of Growth

Essays in Swedish Economic History

Sweden experienced a remarkable economic transformation between the

18th century and the outbreak of World War I. This dissertation consists of

four self-contained papers that uses a quantitative empirical approach to

identify key drivers of this transformation by analyzing the contribution of

the potato to economic growth, the determinants of the early investments

in mass schooling, and how the rollout of the national railroad network

shaped rural and urban growth patterns from the mid-19th century to the

present day. Together, the findings of this dissertation contribute novel

evi-dence on the causal determinants of Sweden’s acceleration in growth and

also shed light on the historical roots of contemporary patterns of regional

and urban development.

Engines of Growth

Essays in Swedish Economic History

© Thor Berger 2016

School of Economics and Management

Department of Economic History

ISBN 978-91-87793-28-8

ISSN 1400-4860

Printed in Sweden by Media-Tryck, Lund University

Lund 2016

Acknowledgements

A dissertation does not get written without one incurring many debts to many

people. Kerstin Enflo has been an outstanding supervisor providing endless

advice, encouragement, and support throughout my time as a PhD student. I

am also very grateful to Martin Henning, who has been my assistant

supervisor providing crucial feedback on my work. Lennart Schön initially

served as my main supervisor but sadly passed away during the writing of

this dissertation, which constituted an unfathomable loss to us all. My work

has been funded by the Crafoord Foundation, the Swedish Research Council,

and the Wallenberg Foundation and their generous financial support is

gratefully acknowledged. At my final seminar, Anders Nilsson and Joan

Rosés provided extremely valuable comments and suggestions that improved

the dissertation, while Martin Dribe and Patrick Svensson guided me through

the process leading up to that seminar as the directors of the PhD programme.

The Department of Economic History in Lund has been an extraordinary

place to work not least due to the diligent management of Anneli Nilsson

Ahlm, Birgit Olsson, Kristin Fransson, Mats Olsson, and Tina Wueggertz, as

well as the many PhD students and post docs that have made these past years

something that I will always remember warmly. I am particularly grateful for

having shared the daily grind with Hana Nielsen, as I could not imagine

anyone better to share the inevitable ups and downs of academic life with. A

final heartfelt thank you is due to my family for their infinite care, love,

patience, and support—without them, none of this would have been possible.

Contents

List of papers

8

Introduction 9

Overview 9

Background: historical perspectives on regional development

11

Why study regions?

11

Sources of regional growth

13

Historical context

20

Sweden’s economic transition prior to World War I

20

Regional growth patterns during industrialization

25

Causes of catch-up: external vs. internal factors

28

Sources 30

Summary of papers

33

Conclusions 37

References 40

List of papers

I.

Berger, Thor. 2016. Geography and growth: Evidence from the

potato’s introduction in pre-industrial Sweden. Unpublished

manuscript.

II.

Andersson, Jens, and Berger, Thor. 2016. Elites and the expansion of

education in 19

_{-century Sweden. Unpublished manuscript.}

III.

Berger, Thor. 2016. Railroads and rural industrialization: Evidence

from a historical policy experiment. Unpublished manuscript.

IV.

Berger, Thor, and Enflo, Kerstin. 2016. Locomotives of local growth:

The short- and long-term impact of railroads in Sweden. Journal of

Introduction

Overview

Sweden was a poor and underdeveloped country in the early 19

_century.

Workers earned real wages well below those of their British, French, or

German counterparts, with most of the population trailing just above

subsistence. Agricultural activities employed more than 90 percent of the

population and less than one in ten lived in a city, most of which would more

accurately have been described as small towns or villages. A century later,

the economy had undergone a fundamental transformation as Sweden

emerged as a modern industrial nation placing it on a trajectory to becoming

one of the richest countries in the world.

A central contribution of this dissertation is the use of historical data

combined with a quantitative empirical approach to identify key drivers of

growth during Sweden’s economic transformation. To isolate the impacts of

these growth determinants the analysis exploits regional data and

quasi-experimental methods, which aligns with a recent methodological shift in

economic history and related fields towards using natural experiments to

identify causal relationships (Diamond and Robinson, 2010). As one of the

remarkable success stories of the 19

_{-century Atlantic economy, Sweden}

constitutes a puzzling example of an initially backward country that managed

to converge with the leading industrial economies within a few generations.

Accounting for this catch-up requires statistical evidence on causal factors

that can explain the acceleration in growth, which arguably constitutes a key

task for Swedish economic historians. Moreover, given the country’s extreme

cultural, ethnic, and religious homogeneity the regional development of the

Swedish economy provides a particularly useful historical laboratory to

cleanly identify such growth determinants, which may carry implications far

beyond the writing of Swedish economic history.

The dissertation consists of an introductory chapter that mainly provides an

overview of Sweden’s economic transition from a variety of perspectives and

four self-contained papers that study different aspects of this transformation

between the mid-18

_{century and the outbreak of World War I. Specifically,}

paper 1 examines the relationship between agricultural productivity increases

and economic growth by quantifying the impacts of the potato’s introduction

in the early 19

_{century on living standards and population growth, which}

shows that the diffusion of the potato was an important catalyst of the

Swedish population explosion that would characterize the economic and

social developments for the remainder of the century. Paper 2 asks how

Sweden managed to maintain a level of human capital and schooling that was

wildly out of proportion to the country’s level of economic development with

a political system that enabled economic and political elites to capture the

political process and block investments in schooling for the masses.

Analyzing the determinants of investments in elementary education,

however, shows that local elites were instrumental in advancing mass

schooling prior to Sweden’s industrial breakthrough, which furthers our

understanding of how the “impoverished sophisticate” came into being. In the

final two papers, I quantify the impact of the railroad on rural development

and its short- and long-term impacts on urban growth respectively. Although

the rollout of the railroad network caused an accelerated pace of population

growth and structural transformation in rural communities traversed by the

state railroads, these gains were confined to the initially most developed areas

thus exacerbating spatial inequalities. As shown in the final paper, the

historical railroads also constituted a shock to the spatial equilibrium of the

urban economy that still shapes the size distribution of Swedish cities today.

A unifying theme of the papers is a regional perspective on the growth

process and in the next section I consequently motivate a regional approach

by arguing that growth in the 19

_{century unfolded along regional lines and}

that the vast income differences that opened up within nations prior to World

War I are challenging to reconcile with explanations that mainly emphasize

national factors. A selective review of the related literature on historical

regional development then provides an overview of the work that this

dissertation advances and builds on. To situate the findings in their wider

historical context I subsequently describe the historical development of the

Swedish economy between the mid-18

_{century and the outbreak of World}

War I. An overview of the data used and a summary of each paper are lastly

followed by a discussion of the broader implications of the results of this

dissertation and potentially fruitful avenues for future research.

Background: historical perspectives on regional

development

A central argument that underpins the analysis in the subsequent papers is

that a regional approach is essential to understand why some areas underwent

rapid economic development while others fell behind in the 19

_century.

_In

this section, I first motivate and highlight the importance of a regional

approach and then selectively review the related literature on historical

regional development, with a particular focus on work that emphasizes the

role of geography, human capital, and investments in transportation

infrastructure to explain spatial patterns of growth.

Why study regions?

A key question for economic historians to answer is why some countries

industrialized early, why some fell behind, and how other initially backward

countries such as Sweden managed to converge with the leading

industrializers. An influential literature by economists, economic historians,

and political scientists that aims to account for the varying growth trajectories

in 19

_{-century Europe typically emphasizes the role of national factors such}

as culture (Weber, 1930; Clark, 2008), human capital (Easterlin, 1981),

institutions (North and Thomas, 1973; Acemoglu et al., 2005), social

capability (Abramovitz, 1986), or a country’s relative backwardness

(Gerschenkron, 1962). National explanations, however, poorly fits the

historical realities of European industrialization and taking the state as the

unit of analysis often introduces a large aggregation bias (Wolf, 2009).

As

highlighted in the work of Pomeranz (2000), for example, comparing

countries of vastly different sizes may conceal the fact that regions within

these countries are at similar levels of development and ”[u]nless state policy

1

_{To be clear, I use the term ”region” to designate any form of subnational unit such as cities,}

counties, states, or parishes throughout the remainder of this introduction.

2 Gerschenkron (1962, p.42) argues that a nation constitutes a more significant unit of

observation than its subdivisions as it represents a “bundle of historical experience”.

However, this obscures the very different regional histories that were merged into nation

states in the 19

_{century (e.g., Wolf, 2009). Pollard (1981, p.187) points out this gap in}

Gerschenkron’s argument by emphasizing that nation states “contain territories at very

different stages of development” and that the degree of backwardness thus varied also

within countries.

is at the center of the story being told, the nation is not a unit that travels very

well” (p.7).

An earlier generation of economic historians was acutely aware of the

regional dimension of growth and industrialization in the 19

_{century. In the}

second paragraph of his magisterial work, Pollard (1981, p.3) emphasizes that

“industrialization in Britain was by no means a single, uninterrupted, and

unitary, still less a nation-wide process”. Consequently, while high wages

and cheap coal may account for why Britain led the world during the

Industrial Revolution (Allen, 2009), it provides less explanatory power as to

why the British cotton industry, the key sector in 19

_-century

industrialization, was almost completely concentrated to Lancashire (Crafts

and Wolf, 2014). Moreover, regional imbalances during industrialization

were by no means confined to the British Isles, as industrial growth

throughout Western Europe was a local rather than national affair (Pollard,

1981, p.41).

Appealing to factors such as country’s culture, institutions, or social

capability is thus insufficient to account for the fact that industrialization did

not adhere to the borders of nation states. Spanish industrialization, for

example, was concentrated to Catalonia and the Basque Country while the

rest of the country progressively deindustrialized (Rosés, 2003). In Italy, the

Northern “Industrial Triangle”—Liguria, Lombardy, and Piedmont—

modernized as the Mezzogiorno fell behind (Felice, 2011). At the eve of

Swedish industrialization, regional differences in income per capita similarly

varied by a factor 3.4 between the richest and poorest region, considerably

larger than the gap in terms of per capita income between Sweden vis-à-vis

Germany (1.3), France (1.5), and the United Kingdom (2.3).

More broadly,

while national factors such as institutions may explain average differences in

income between countries, they sit uneasily with the fact that income

differences between European countries were smaller than differences within

countries in the 19

_{century (Caruana-Galizia, 2015). Explaining these}

regional disparities is of crucial importance even if one is mainly concerned

with national differences in income, since a country’s prosperity in the end is

nothing more than the sum of its regions. Moreover, understanding why these

regional gaps opened up in the 19

_{century is not only of historical interest,}

3 My own calculations based on data on GRP per capita in 1860 measured in 1910/12 SEK

from Henning et al. (2011) and national GDP per capita data for the same year measured in

1990GK$ from The Maddison Project

(http://www.ggdc.net/maddison/maddison-project/home.htm).

but also carry important policy implications since these disparities in many

cases persist until the present day.

A complementary motivation for a regional approach is methodological in

nature. Although a large body of work in the vein of Barro (1991) has strived

to identify the factors that shape global income inequalities, such an analysis

is often plagued by measurement issues and unobserved country-level

factors. As a response to this challenge, empirical work by economists and

economic historians has increasingly shifted toward exploiting

within-country variation to identify potential growth determinants. Comparing

regions located within the same country is often eased by the fact that

statistical agencies collect consistently defined regional data and further

enables us to hold many factors that are challenging to measure constant,

which reduces concerns that the observed relationships in the data are

spurious. A recent literature indeed highlights the benefits of using a regional

approach to identify the role of, for example, institutions, human capital, or

social capital (e.g., Dell, 2010; Acemoglu et al., 2011; Nunn and

Wantchekon, 2011; Gennaioli et al., 2012). Although a regional approach is

valuable in itself to understand historical patterns of growth and

industrialization it thus also offers a useful setting to gain a more general

understanding of the causal determinants of growth.

Sources of regional growth

First and second nature geography

A large body of work emphasizes the role of geography as a fundamental

determinant of economic development. One strand of this literature stresses

the role of physical geography, by arguing that the distribution of “first

nature” advantages such as arable land, climate, and rainfall directly affects

productivity or determines the appropriate set of technologies (e.g., Sachs and

Warner, 1995; Diamond, 1999; Jones, 2003; Olsson and Hibbs, 2005; Allen,

2009). Alternatively, the distribution of such first nature advantages may

indirectly affect income levels by, for example, shaping institutional

development or social capital (e.g., Sokoloff and Engerman, 2000; Easterly

and Levine, 2002; Acemoglu et al., 2002, 2005; Nunn and Wantchekon,

2011; Nunn and Puga, 2012).

Agricultural potential and opportunities for land- and water-based

transportation were a dominant force in shaping the spatial distribution of

economic activity throughout the pre-industrial era thus highlighting the role

of first nature advantages (Bairoch, 1988; Ashraf and Galor, 2011; Bosker

and Buringh, 2015). For example, after the potato’s introduction in the wake

of the 16

_{-century Columbian Exchange, areas endowed with land suitable}

for potato cultivation in the Old World experienced substantially more rapid

population growth and urbanization (Nunn and Qian, 2011). Although

transportation costs eventually declined, the early stages of the Industrial

Revolution took place in a context of high transport costs thus making the

distribution of first nature advantages key in understanding the spatial

patterns of early industrial growth. Coal was king during the Industrial

Revolution and as Pollard (1981, p.4) concluded: “the map of the British

industrial revolution, it is well known, is simply the map of coalfields”,

which is further underlined by the recent findings of Fernihough and

O’Rourke (2014) that European cities located in proximity to coal fields saw

differentially more rapid growth after coal began to be used in industrial

processes. Although much of the historical literature has emphasized the role

of such first nature advantages, however, an alternative force in shaping the

distribution of economic activity is “second nature” geography.

A growing body of work emphasizes the role of second nature geography

factors that accrue from the gravitational pull of economic activity itself. At

least since Marshall (1890), it has been known that one central motivation for

concentrations of economic activity is the existence of agglomeration

economies that are derived from an improved “matching, sharing, and

learning” in areas with a higher density of economic activity (Duranton and

Puga, 2004). As firms and workers concentrate it allows them to save on

transaction and transportation costs, it improves the matching between

employers and workers by creating thicker labor markets (Kim, 2006), and it

eases the spread of new ideas or innovations due to more frequent

interactions (Simon and Nardinelli, 1996). At the heart of these explanations,

however, is a coordination problem: consumers and producers tend to prefer

locations that offer larger markets and these larger markets tend to be exactly

those locations where other consumers and producers choose to locate.

Among any number of identical locations, the one that for whatever reason

attracts a critical mass of economic activity may thus draw in even more

activity thus giving rise to the potential for multiple equilibria (Krugman,

1991a,b,c). An interesting implication is that if agglomeration economies are

sufficiently strong, even minor historical accidents that coordinate economic

activity to a certain location may permanently affect the spatial distribution of

economic activity, which implies that patterns of economic activity may be

explained as much by chance as by geographic fundamentals.

A central question for economists, economic historians, and geographers is

thus whether the spatial distribution of economic activity is determined by

first or second nature geography. A formal analysis of their respective role is

provided in Bleakley and Lin (2012) that examine urban formations in the

19

_{-century United States that arose around sites along rivers that were}

suitable for portage. By the early 20

_{century, the need to haul water crafts}

and cargo over land to avoid water obstacles had become an economically

irrelevant activity due to the spread of the railroad network, yet the cities that

had formed around portage sites persist until the present day. A persistence of

economic activity around portage sites indeed suggests an important role for

second nature geography and path dependence in shaping the geographic

distribution of economic activity. Yet, others find little evidence of path

dependence playing a central role in explaining the spatial pattern of

economic activity. Davis and Weinstein (2002, 2008), for example, examine

the impacts of the devastating bombing campaigns of Japanese cities during

World War II. Evidence that these extreme reductions in city populations had

no long-run effect on cities’ industrial structure or population instead

suggests that first nature geography is the most relevant explanation for the

spatial structure of the economy.

More broadly, a growing historical

literature examines the respective role of both first and second nature

geography in shaping the spatial distribution of economic activity from the

pre-industrial era to the 20

_{century suggesting that while both factors}

mattered, second nature factors became relatively more important as

industrialization progressed (e.g., Kim, 1995; Rosés, 2003; Crafts and

Mulatu, 2005; Wolf, 2007; Klein and Crafts, 2011; Crafts and Wolf, 2014;

Bosker and Buringh, 2015; Hanlon, 2015).

4 Although much of this literature has exploited natural experiments, which may have limited

external validity, recent contributions have also evaluated the feasibility of “big push”

policies. Kline and Moretti (2013), for example, examine the long-run impacts of the

Tennessee Valley Authority, one of the largest regional policy initiatives in the history of

the United States, showing that it had persistent effects on county-level industrial activity,

which is taken as evidence for the existence of multiple equilibria.

5 A series of papers have used the historical division and reunification of Germany as a natural

experiment to identify the role of market access and second nature geography forces.

Redding and Sturm (2008) show that German cities that experienced a larger loss of market

access after the division of Germany saw a reduction in growth, Ahlfeldt et al. (2015)

provides evidence on the role of agglomeration economies from the division of Berlin,

while Redding et al. (2011) study the relocation of Germany’s main airhub from Berlin to

Frankfurt showing that there is little evidence of a shift back to Berlin after reunification,

which they take as evidence of multiple equilibria in industry location.

Human capital, local institutions, and the spread of mass schooling

Human capital is at the center of modern growth theories (Lucas, 1988;

Romer, 1990; Mankiw et al., 1992; Galor, 2005), yet traditional accounts of

the Industrial Revolution typically downplay the role of human capital due to

Britain’s inadequate system of formal schooling. As famously summarized

by Mokyr (1992, p.240): “If England led the rest of the world in the

Industrial Revolution, it was despite, not because of her formal education

system.” A minor role for human capital during early industrialization chimes

well with views of technological change as “deskilling” and is consistent

with the argument of Acemoglu (2002, p.8) that “the idea that technological

advances favor more skilled workers is a twentieth-century phenomenon”.

A recent body of work, however, documents the crucial role of human

capital and schooling in follower countries building on the observation by

Cipolla (1969, p.87) that “more literate countries were the first to import the

Industrial Revolution”. Becker et al. (2011), for example, show that Prussian

regions that early on invested in schooling were more successful during

late-19

_{century industrialization, Franck and Galor (2015) show that}

industrializing French regions had higher literacy rates and invested in more

schools, while Squicciarini and Voigtländer (2015) presents evidence that the

spatial distribution of scientific elites in early modern France shaped the

adoption of advanced technologies during the Industrial Revolution, which

led to higher growth, productivity, and wages in modern sectors. Mounting

evidence that early investments in human capital seemingly caused industrial

development has sparked a renewed interest in identifying factors that can

account for the uneven spread of mass schooling in the 19

_century.

A weak link between national schooling laws and educational

performance, however, has led such research efforts to focus particularly on

understanding the uneven spread of schooling within countries and the local

determinants of educational investments.

In particular, Lindert (2004)

advocates such a bottom-up perspective and emphasizes that decentralized

control over taxes and schooling permitted regions that demanded education

to forge ahead unfettered from conflicts between national elites. Go and

6 Although 19

_{-century educational leaders such as Prussia early on enacted compulsory}

schooling laws, they were seemingly ineffective in raising enrollments in other countries,

or were clearly preceded by an expansion of schooling. In France, for example, enrollments

increased well before Jules Ferry’s Laic Laws of the 1880s, whereas the 1857 Moyano Act

in Spain and similarly early compulsory schooling laws in other Southern European

countries such as Greece and Portugal led to disappointing enrollment rates in the latter

Lindert (2010) provide empirical evidence that the autonomy of local

governments and a widely spread political voice indeed can account for why

the Northern United States led the world in terms of primary schooling

around the mid-19

_{century. An emphasis on the role of political voice is also}

found in Galor et al. (2009) that theoretically and empirically study the role

of landed elites in the United States, concluding that high land inequality was

an impediment to educational investments during the high school-movement.

Evidence on similar mechanisms that emphasize the negative influence of

elite control and a limited political voice has been found in colonial Korea,

Prussia, and Spain (Go and Park, 2012; Beltran Tapia and

Martínez-Galarraga, 2015; Cinnirella and Hornung, 2016). Another strand of the

literature, however, has stressed that political inequality was not necessarily a

barrier to the spread of mass schooling since elites often supported its spread

motivated by ethnic favoritism or social control motives (Cvrcek and Zajicek,

2013; Gao, 2015; Shammas, 2015), while a third body of work downplays

the role of political factors and instead has stressed the key role of religion as

a central driver of differences in human capital development, particularly

emphasizing the role of the Reformation and the uneven spread of

Protestantism (e.g., Becker and Woessmann, 2009).

As highlighted by the varying findings of this growing literature, the

determinants of educational expansions are likely to vary considerably both

across countries and within individual countries over time as they were

shaped by interactions between a plethora of national and local factors. For

example, although decentralization may be beneficial at early stages of

school development, it may have adverse effects on investments in schooling

as the system matures (Goldin and Katz, 2009; Capelli, 2015). Similarly,

whether a limited political voice and concentration of power into the hands of

a few had adverse effects on the development of schooling critically hinges

on whether economic and political elites perceived and educated populace as

an economic opportunity or political threat (Kaestle, 1976; 1983). Furthering

our understanding of the factors that determined the uneven spread of mass

schooling both across and within countries in the 19

_{century thus requires}

careful empirical studies of individual countries during various stages of their

development of mass schooling to complement the existing cross-country

work of a more comparative nature (e.g., Lindert, 2004).

Local economic development and the “transportation revolution”

An increased role of second nature geography over the 19

_{century was}

intimately associated with the “transportation revolution”. Major investments

in canal construction, the building and improvements of roads, the

construction of railroad networks, and the advent of the steam ship led to a

precipitous decline in transportation costs (O’Rourke and Williamson, 1999,

pp.33-36). Lower transport costs in turn meant access to wider markets,

cheaper raw materials, and higher rates of factor mobility, which was the

foundation for an industrial system in which factories churned out

standardized goods for mass markets. As forcefully summarized by Pollard

(1981, p.123): “Industrialization means mass production, and mass

production means access to mass markets. An industrial revolution, it has

always been known, must be associated with a transport revolution.”

Above all, the railroad was the key innovation that epitomized the

transportation revolution to contemporary observers and its role loom large in

accounts of 19

_{-century development.}

_{Yet, the early cliometric literature}

that evaluated the contribution of the railroad to the national economy found

little support for its presumed “indispensability” for economic development

(Fogel, 1964).

_{More recently, economic historians have turned to quantify}

the impacts of the railroad on local economic development rooted in the

belief that the coming of the railroad had a variety of neglected economic

impacts such as increasing the mobility of capital and labor, promoting

economies of scale in manufacturing, and easing the spread of new ideas and

knowledge. A central empirical challenge in assessing the impact of

infrastructure investments on local economic development, however, is to

identify its causal effect since railroads were often built to connect already

rapidly growing places (Redding and Turner, 2014). In the context of the

7 Rostow (1960, p.55), for example, argues that “[t]he introduction of the railroad has been

historically the most powerful single initiator of take-offs” and Landes (1969, p.153)

similarly states that “by the 1840s railroad construction was the most important single

stimulus to industrial growth in Western Europe”.

8 Cliometric work has largely focused on the estimating the “social savings” of the railroad,

which in its simplest form corresponds to the cost savings of transporting goods using the

railroad relative to alternative transport modes. See O’Brien (1977, 1983), Crafts (2004),

and Leunig (2010) for an overview of this literature. More recently, Donaldson and

Hornbeck (2016) revisited Fogel’s analysis of the importance of the railroad to American

economic development using modern Geographic Information Systems (GIS) data and

trade theory to derive a reduced-form relationship between market access and land values

to show that removing all railroads in 1890 would lead to an annual reduction of GNP of

about 3.2 percent, which is of a similar magnitude as the estimates by Fogel (1964).

Antebellum United States, Fishlow (1965, p.203) famously highlighted this

challenge by arguing that:

A key issue, however, is whether such railroad influence was primarily

exogenous or endogenous, whether railroads first set in motion the forces

culminating in [...] economic development [...] or whether arising in response

to profitable situations, they played a more passive role.

A growing body of evidence suggests that railroads in fact “set in motion”

a wide variety of forces that in crucial ways shaped patterns of local

economic development in the 19

_{century. Recent evidence from the United}

States, for example, links the diffusion of railroads to agricultural

improvements (Atack and Margo, 2011), the transition from the artisan shop

to the factory (Atack et al., 2008), and urbanization (Atack et al., 2010).

Evidence from Central Europe similarly suggest that the spread of the

railroads accelerated city growth and improved market integration (Keller

and Shiue, 2008; Hornung, 2015), while there is growing evidence that

investments in transportation infrastructure significantly raised growth also in

contemporary developing countries. Donaldson (2015), for example, shows

that railroads constructed under the British Raj decreased trade costs,

eliminated the adverse effects of local productivity shocks, and raised real

incomes in colonial India, while Jedwab and Moradi (2016) document the

lingering positive impacts of African colonial railroads on agricultural

production and city growth.

However, although transportation improvements are generally believed to

have benign effects in connected areas, theoretical developments in the New

Economic Geography (NEG) that emphasize the interaction between

increasing returns to scale and falling transportation costs suggest that

peripheral areas may be impoverished as transport costs decrease (Krugman,

1991c). Krugman (1991c, p.97) himself, for example, argues that “railroads

and steamships led to the deindustrialization of the periphery” in the 19

century. Similarly, while Donaldson (2015) documents the positive impacts

for areas connected to the railroad network in colonial India, he also finds

that a connection hurt neighboring regions. In Meiji Japan, the construction

of the railroad network led to a concentration of manufacturing activity to a

few regions (Tang, 2014), while the German railroad boom similarly

contributed to a deindustrialization of peripheral areas (Gutberlet, 2014).

Thus, while the expanding literature on the impacts of the transportation

revolution suggests that the coming of the railroad was deeply intertwined

with rapid agricultural development, industrialization, and urbanization, the

more recent contributions also stress that the overall impact on regional

inequality may be more ambiguous.

Historical context

In order to situate the four individual papers in their historical context, this

section provides a brief overview of Sweden’s economic transition prior to

World War I, the uneven patterns of regional growth during industrialization,

and the respective role of external and internal factors in accounting for

Swedish catch-up with the industrial leaders.

Sweden’s economic transition prior to World War I

Figure 1.1 shows GDP per capita, population, and real wages in Sweden

between 1750 and the outbreak of World War I. A stagnating economy in the

latter half of the 18

_{century is evident both from the GDP figures and from}

the fact that real wages declined to their lowest level since the early modern

era during the Napoleonic Wars (Söderberg, 2010). Over the first half of the

19

_{century, however, the Swedish economy for the first time began to}

exhibit sustained growth with per capita incomes increasing by some 0.3-0.4

percent on an annual basis (Schön, 2010, p.49). At the same time, population

growth accelerated and between 1750 and 1850 Sweden’s population doubled

in size, with a particular expansion among the rural lower classes. Although

there is little indication of any secular changes in real incomes for the masses

prior to the mid-19

_{century, most indicators of health improved substantially}

well before then (Sandberg and Steckel, 1997; Bengtsson and Dribe, 2002).

An influential explanation for the empirical puzzle of simultaneous

improvements in health, population growth, and stagnating incomes is the

contemporary poet Esais Tegnér’s emphasis on “the peace, the potato, and

the vaccine”, which is still echoed in the writings of Swedish economic

historians (e.g., Sandberg and Steckel, 1988; Myrdal and Morell, 2011;

Schön, 2010).

Figure 1.1 Swedish GDP per capita, population, and real wages, 1750-1914.

Notes: GDP per capita is based on data in Krantz and Schön (2007), population is drawn from LU-MADD (http://www.ekh.lu.se/en/research/economic-history-data/lu-madd), and real wages are based on the Stockholm series constructed by Söderberg (2010).

Over the latter half of the 19

_{century, Sweden exhibited the most rapid}

growth among Western European countries: between 1870 and 1910, GDP

per capita grew at an annual rate of 1.9 percent (see Table 1.1). Growth was

fueled by capital accumulation and total factor productivity (TFP) increases

that outpaced those of all countries for which data is available, with an annual

growth of the capital stock and TFP of 3.4 and 1.5 percent respectively.

Although real wages had stagnated during the first half of the century, the

benefits of growth were spread more widely when industrialization

accelerated in the latter half of the 19

_{century, which is reflected in secular}

wage increases (see Figure 1.1). At the eve of World War I, real wages had

increased six fold relative to the trough a century before, which was mirrored

in rapid wage convergence with many of the leading industrial economies

(Williamson, 1995; Prado, 2010).

Table 1.1 Swedish growth in a Western European perspective, 1870-1910.

Notes: This table reports annual growth rates of GDP per capita, the capital stock, and TFP between 1870 and 1910, based on data from Krantz and Schön (2007) and Carreras and Josephson (2010).

GDP p.c. (%) Capital stock (%) TFP (%)

Sweden 1.90

3.43

1.46

Denmark 1.57

3.29

1.32

France 1.45

1.41

1.19

Germany 1.72

3.12

1.24

Italy 0.92

2.67

0.48

Netherlands 0.89 3.14 0.54

Spain 1.23

1.82

1.12

United Kingdom

0.97

2.13

0.48

Sweden’s economic transformation was driven both by advances in

productivity within sectors and a structural shift from agricultural to

industrial activities. Agriculture underwent a veritable revolution between the

mid-18

_{and mid-19}

_{century that involved fundamental changes in}

agricultural practices, the ownership of land, and technological advances

(Gadd, 2000). A key factor in accounting for these changes is the Enclosure

Movement that gained ground around the mid-18

_{century: in 1749, the Field}

Consolidation Act promoted a consolidation of landholdings and the

Redistribution Act of 1827 gave each peasant the right to demand that his

land be enclosed. Evidence from farm-level production accounts suggests

that the breaking up of the village system and the shift from open-field

agriculture to enclosures led to substantial increases in productivity (Olsson

and Svensson, 2010), which is consistent with a steadily increasing

agricultural productivity from the early 19

_{century and onwards as the}

enclosures spread throughout the country (Myrdal and Morell, 2011, p.157).

A central lever for these productivity advances was the individualization of

production decisions that the enclosures enabled, which promoted the

adoption of new crops, technologies, and tools (Gadd, 1983). An introduction

of new crop rotations and a reduction or elimination of the fallow, the

widespread adoption of iron ploughs, an increased use of draught animals,

investments in drainage and soil improvements, as well as the replacement of

sickles with scythes all significantly contributed to growth (Myrdal and

Morell, 2011, p.145f).

Among the many important changes was the

introduction of the potato as a field crop, which Heckscher (1954, p.150)

terms as “revolutionary”. Together, these changes permitted considerable

land reclamation, which led to the total cultivated acreage doubling in size

over the first half of the 19

_{century (Hannerberg, 1971). Many of these}

changes required significant investments and the spread of mortgage

associations and savings banks in the countryside constituted new crucial

sources of capital to finance the agricultural revolution (Nygren, 1985). As

the Enclosure Movement reached its peak in the mid-19

_{century, the}

productive capacity of the agricultural sector had increased substantially:

while Sweden had imported around a tenth of the grain used for domestic

consumption in the early 19

_{century, the country had turned to a net}

exporter five decades later (Schön, 2010, p.59).

A transformation of agriculture created the preconditions for early

industrialization by creating a demand for manufactured goods and by

increasing the supply of labor. Over the first half of the 19

_{century, growing}

incomes and an increased supply of landless agricultural laborers fueled

proto-industrial advances in the form of cottage industries in the countryside

(Schön, 2010, p.76). At the same time, industrial advances took place in

textiles where factory production gradually appeared to satiate the growing

demand for manufactured cloth (Schön, 1979). As incomes increased for an

increasingly larger share of the population, the manufacturing of simpler

products for private consumption such as candles, matches, and soap also

expanded (Schön, 2010, p.155). However, Sweden considerably lagged

behind the rest of Western Europe in terms of industrial development (see

Table 1.2), which mirrored the fact that industrial advances during the first

half of the century had been relatively limited in scope and had failed to

ignite a technological transformation of the manufacturing sector that is

evident from the limited use of steam power that trailed far behind other

European countries (Kander et al., 2014, p.184).

9

_{Nilsson et al. (1999) also emphasize the bureaucratic nature of the enclosure process that}

required agreements, maps, and petitions to be signed, read, and understood as well as the

increasingly impersonal market transactions as two factors that increased the demand and

supply of literacy among peasants in the wake of the Enclosure Movement.

Table 1.2 Swedish industrialization in a European perspective, 1860-1913.

Notes: This table reports average levels of industrialization indexed to the United Kingdom in 1900 = 100 based on data in Bairoch (1982, pp.294, 330) and Broadberry et al. (2010, p.70). Central and Eastern Europe is an unweighted average of Austria-Hungary, Bulgaria, Germany, Romania, Russia, Serbia, and Switzerland; North-western Europe is an unweighted average of Belgium, Denmark, Finland, the Netherlands, Norway, and the United Kingdom; while Southern Europe is an unweighted average of France, Greece, Italy, Portugal, and Spain.

Year 1860 1880 1900 1913

Sweden

15 24 41 67

Central

and

Eastern

Europe 11 16 26 37

North-western

Europe

23 31 40 53

Southern

Europe

11 14 19 26

European

average

17 23 33 45

Modern economic growth had its breakthrough in Sweden in the latter half

of the 19

_{century as investments in infrastructure and industrialization}

emerged as new engines of growth. The transformation of agriculture and the

early industrial advances had increased the need for improving the

transportation system and an early wave of infrastructure investments took

place between the 1780s and 1820s, with the majority of state funds being

allotted to canals that aimed to connect the major lakes in the interior to the

coast (Westlund, 1998). While an expansion of the canal system also

triggered complementary investments in roads to supplement the waterways,

the construction of the railroad network beginning in the 1850s was by far the

most ambitious infrastructure project of the century: investments in the

transport sector, primarily expenditure on railroads, constituted roughly a

third of the total investments in the economy in the latter half of the 19

century (Schön, 2010, p.119).

Manufacturing growth expanded to encompass a broader set of industrial

activities from mid-century, fueled by a growing foreign demand for Swedish

natural resources and transportation improvements. Although traditional

exports such as iron remained important, it was particularly two new exports

that gained ground: oats and timber. Exports of oats experienced a sustained

boom in the 1850s and 1860s owing to the growing demand from the

expanding British industrial cities, while the demand for Swedish timber led

to an export-led boom in the peripheral northern regions that was enabled by

the clearing of the float ways and the adoption of steam saws. As described

by Heckscher (1954, p.226), the timber industry expanded in “a

gold-rush-like experience to which there was no counterpart in earlier Swedish history”

during the two major upswings of the 1850s and the 1870s. Although timber

exports decreased in importance by the turn of the century, a shift towards

paper and pulp production ensured the importance of the “green gold” well

into the 20

_{century, as Sweden became the world’s leading pulp exporter}

(Heckscher, 1954, p.228).

Among Swedish economic historians there is a consensus that the 1870s

constitute somewhat of a watershed in the industrialization process as the

preconditions for a more widespread industrial breakthrough had been

created (Gårdlund, 1942; Heckscher, 1954; Schön, 2010). Industrial growth

indeed accelerated in new directions during the international boom of the

1870s and over the coming decades the Swedish manufacturing industry

became increasingly sophisticated. Growth shifted from older industries such

as ironworks and saw mills toward chemical industry, consumer goods, and

mechanical engineering that manifested Sweden’s shifting comparative

advantage (Jörberg, 1961). As the century drew to a close, the Swedish

industry experienced a final growth spurt that unquestionably transformed

Sweden into a modern industrial economy, which is evident from the fact that

many of the technologically sophisticated firms that would constitute the

basis of Swedish industry in the 20

_{century such as L.M. Ericsson and SKF}

were formed around domestic and foreign technological innovations around

the turn of the century. As evident from Table 1.2, this accelerated pace of

industrialization firmly established Sweden as part of the North-western

European industrial core by the outbreak of World War I—within a few

generations the country had been transformed from an industrial laggard to

one of Europe’s leading industrial nations.

Regional growth patterns during industrialization

Although the transformation of the Swedish economy was one of gradually

accelerating growth, it was also characterized by substantial shifts in the

regional patterns of development (Söderberg, 1984). Writing in the 1960s,

Jeffrey Williamson hypothesized that regional inequalities tend to increase

during the early stages of the industrialization process, while they tend to

decrease as an economy develops when the extension of infrastructure

networks, migration, and the spread of industrialization serve to reduce

spatial disparities (Williamson, 1965).

Figure 1.2 Regional inequality during Swedish industrialization, 1803-1914.

Notes: This figure shows the unweighted coefficient of variation of agricultural day laborer’s nominal wages across Swedish counties between 1803 and 1913 based on data from Jörberg (1972). Also shown is a fitted local polynomial regression estimate.

Figure 1.2 revisits the Williamson hypothesis in the context of 19

_-century

Sweden, by examining the variation in nominal county-level wages for

Swedish agricultural day laborers between 1803 and 1914 drawn from

Jörberg (1972). In the latter half of the 18th century, wages decreased in

several regions and the subsequent catch-up over the first half of the 19th

century was unevenly spread as wages stagnated or even decreased in many

areas (Jörberg, 1987). As shown in Figure 1.2, these uneven growth patterns

are reflected in rising regional inequalities from the early 19

_{century until}

the 1870s, which lends broad support to the Williamson hypothesis. Rising

regional inequalities have been attributed to the uneven agricultural

transformation, localized industrial advances, and the greater dynamism of

the western and southern parts of the country (Schön, 1979; Söderberg, 1984;

Bengtsson, 1990), while an unbalanced regional development also reflected

an underdeveloped transportation system characterized by high costs of

moving both goods and people.

Although Sweden enjoys ample opportunities for water transportation due

to its extensive coastline, overland transport constituted a thorny problem in

most areas in the 18

_{and early 19}

_{century. Prior to the railroad era, land}

transportation was confined to using horses and wagons on dirt roads during

the summer months and “[t]he lack of a developed highway system was

acutely embarrassing in a country as extensive and sparsely populated as

Sweden” (Heckscher, 1954, p.240). Transport conditions, however, improved

in the wintertime due to the fact that the frozen ground and ice-covered lakes

and watercourses—the “winter roads”—eased land transport (Westlund,

1998, p.67). Yet, as a consequence of an underdeveloped transportation

system, urbanization rates in Sweden lagged behind the rest of the Western

European industrializers: about one in ten Swedes lived in a city by the

mid-19

_{century and most parts of the country lacked any significant urban}

agglomerations altogether.

Even though urbanization accelerated in the

late-19

_{century as manufacturing increasingly relocated to cities, Sweden’s}

industrialization remained a largely rural phenomenon well into the 20

century (Söderberg, 1984; Berger et al., 2012).

Contemporary observers were acutely aware that the absence of

communications had retarded economic development in many areas, which

provided the impetus to the regional policy ambitions of the state railroads.

State planners such as Nils Ericson forcefully argued that the trunk lines of

the network should to the largest extent possible be routed through

disadvantaged areas that lacked other means of transport thus avoiding

existing transportation networks and historically important cities (Heckscher,

1954). As the railroad network was rolled out, many economically marginal

areas along the routes of the trunk lines indeed gained access to the emerging

network, which has led economic historians such as Westlund (1998, p.74) to

conclude that:

The trunk lines, quite simply, were that epoch’s great instrument of regional

policy for spreading industrialization and economic development to new

regions.

Against the backdrop of railroad diffusion, regional price differentials

narrowed considerably, real wages converged both across and within

occupational groups and regions, and the declines in regional income

inequality that started in the last decades of the 19

_{century was the}

beginning of a trend that would be sustained for the next 100 years (Jörberg,

10 Swedish cities were furthermore remarkably small: while the capital Stockholm housed

some 75,000 inhabitants, only two other cities (Gothenburg and Karlskrona) had more than

10,000 inhabitants in 1800 (Nilsson, 1992). Thorburn (2000, pp.148-149) argues that the

high transport costs that characterized the pre-rail era set a ceiling to city size at about

10,000 inhabitants in most inland regions.

1972; Lundh et al., 2005; Henning et al., 2011).

A central explanation for

the reduction in regional inequality involves the fortuitous location of natural

resources in the peripheral northern regions that allowed them to take part in

industrialization and the comparatively high rates of emigration and

inter-county migration that served to reduce regional differences in productivity

and wages (Söderberg, 1984; Enflo et al., 2014; Enflo and Rosés, 2015). In

these respects, Sweden was relatively unique in a European comparison as

regional inequalities typically increased in other countries during the decades

prior to World War I (Rosés et al., 2010; Combes et al., 2011;

Caruana-Galizia, 2015). Together, the sustained reductions in regional inequality that

began as the century drew to a close thus lends strong support to

Williamson’s hypothesis that spatial disparities tend to decrease as a country

develops (see Figure 1.2).

Causes of catch-up: external vs. internal factors

A longstanding debate among Swedish economic historians concerns the

respective role of external and internal factors in accounting for Sweden’s

rapid industrialization in the latter half of the 19

_{century. Consequently, two}

broad explanations exist that emphasize the role of foreign influences and

domestic dynamics respectively (Schön, 1997).

An influential body of work emphasizes the role of foreign demand for

Swedish natural resources, which ultimately created the preconditions for

industrialization. Summarizing this “export-led model” of Swedish

industrialization, Jörberg (1973, p.439) argues that “Sweden’s industrial

development was in high degree a process of adaptation to events outside the

country’s frontiers [...] only to a lesser extent was it an independent process

of economic expansion”.

_{Along similar lines, O’Rourke and Williamson}

(1995a,b) stress international trade and Heckscher-Ohlin forces in accounting

for Swedish catch-up in the late-19

_{century Atlantic economy. As the pace}

of growth accelerated around the mid-19

_{century, foreign trade indeed had}

11

_{However, the level of market integration was still sufficiently low for the localized harvest}

failures of 1867-68 to have catastrophic consequences, which ignited the first wave of mass

emigration to America from the hardest hit areas. Emigration was extensive from these

regions over the coming decades, with the rate of emigration varying inversely with the US

and Swedish business cycles (Bohlin and Eurenius, 2010).

12

_{Heckscher (1954, p.209) similarly notes that “the stream of influences to Sweden}

considerably outweighed those from Sweden. Generally speaking, Sweden did not cease

being on the receiving end until around 1910”.

its major breakthrough: between 1850 and 1870, export volumes grew at an

annual average of 7 percent, which was a tenfold increase relative to the

preceding 50 years (Schön, 2010, pp.119-120). At the eve of World War I,

the volume of exports had increased by a factor of twenty-six, which

underlines the increasing importance of trade and the international orientation

of the Swedish economy (Heckscher, 1954, p.212).

A focus on exports and trade implicitly puts much weight on a series of

liberal policy reforms, which arguably were crucial to allow Sweden to

exploit the economic opportunities of the First Globalization. In 1846, the

Guild Ordinance was revoked and two years later, the Companies Act

introduced limited liability for company owners, which was to prove “vital to

the process of industrialization” (Schön, 2010, p.131). Finally, the 1864

Freedom to Trade Act enabled anyone of legal age to establish a company

thus firmly reducing the barriers to entry. A liberalization of domestic market

institutions was also mirrored in a number of successive trade reforms in the

1850s and 1860s: free trade was introduced in agricultural goods, import

tariffs were considerably reduced on manufactured goods, all prohibitions on

exports and imports were removed, and as Sweden joined the Cobden

Chevalier treaty in 1865 the breakthrough of virtually free trade had taken

place, which was further eased by the adoption of the gold standard less than

a decade later (Heckscher, 1954, p.237). As a result, Sandberg (1979, p.226),

for example, argues that “[i]nstitutional, cultural, social, and legal obstacles

to the introduction of new technologies or the efficient use of resources did

not exist” by the middle of the 19

_{century. As noted by Schön (2010,}

pp.125-127), however, history is replete of examples of countries that have

undergone pervasive institutional reforms or a short-lived export boom of

primary products that have failed to set in motion a process of sustained

economic development.

A second strand of explanations therefore instead put more emphasis on

the role of internal factors such as the early expansion of industrial activities

in response to the agricultural transformation or the disproportionate

pre-industrial accumulation of human capital as factors that facilitated Swedish

industrialization. Schön (1979), for example, documents the expansion of

textile production in the first half of the 19

_{century that arose as growing}

incomes for wealthier landowners and later broader parts of the peasant

population led to a rising demand for factory-made cloth. An expansion of

domestic crafts and the rural cottage industry similarly constitute evidence of

early industrial advances, well before the upswing in exports around

mid-century (Schön, 1982; Magnusson and Isacson, 1982; Gadd, 1991). A related

explanation puts more weight on the disproportionate accumulation of human

capital prior to industrialization. Cipolla (1969), for example, estimates that

Swedish literacy was about 90 percent in 1850, which led Sandberg (1979) to

term Sweden the “impoverished sophisticate” and attribute the acceleration of

Swedish growth in the latter half of the 19

_{century to the high initial stock of}

human capital.

Based on these observations, Ljungberg and Schön (2013)

summarize the “domestic market model” by emphasizing that the agricultural

transformation, the equal distribution of land, and the high levels of literacy

and human capital created the fundamental preconditions for successful

catch-up in the latter half of the 19

_century.

To sum up, while traditional accounts mainly have emphasized the role of

external factors in accounting for Sweden’s catch-up with the industrial

leaders, the results from more recent research has shifted towards

emphasizing internal factors. In this light, Sweden’s rapid industrialization in

the late-19

_{century did not constitute a radical break with the past, but rather}

had its roots in the agricultural transformation and the early industrial

advances of the first half of the century.

Sources

Each paper of this dissertation draws on newly collected data from a variety

of historical sources that I combine with information from a range of GIS

databases. This section briefly describes these sources, while additional data

used is described in more detail in each individual paper.

Swedish historical statistics are unique in an international perspective and

are widely considered to be of high quality owing to the establishment in the

mid-18

_{century of a central statistical agency in the shape of the Tabular}

Commission.

_{The Tabular Commission maintained the earliest systematic}

Swedish population statistics that began to be constructed in 1749, which

13 Consequently, while Sweden was an economically backward country in the early 19

century its level of development was higher if one prefers alternative measures that put

more weight on human capital. According to the Human Development Index (HDI) derived

by Crafts (2002), for example, Sweden stands out as a forerunner in the early 19

_century:

in 1820, Sweden had a HDI of 0.40 thus putting it ahead of countries such as France (0.30),

Germany (0.34), and the United Kingdom (0.38) (see Pamuk and van Zanden, 2010,

p.231).

14 Sandberg (1979, p.226), for example, muses over the fact that “one of the very poorest

countries in Europe had by far the most elaborate system of national recordkeeping.“

were based on two pre-printed forms (the “Population” and “Mortality” form

respectively) that the clergy in each parish used to report the age-, sex-, and

occupational composition of the parish population as well as to record

demographical events such as fertility, mortality, migration, and nuptiality.

In addition, these forms also contained a number of additional tables where

the clergy reported, for example, the number of individuals who received

poor relief or the number of children enrolled in schools. Although most

variables are available at five-year intervals, demographical events were

often reported at an annual basis in the 19

_{century. Yet, it was not}

uncommon that forms were destroyed in fires or during transport from the

parishes to the Tabular Commission in Stockholm, or that the clergy simply

did not receive the forms, which means that data is not necessarily available

for all parishes and years. However, despite these potential drawbacks the

information provided by the Tabular Commission is unique in a European

context in that it allows for a detailed and long-term study of processes of

population growth and local economic development already from the

mid-18

_{century. Paper 1 uses a balanced set of parishes that reported information}

to the Tabular Commission between 1750 and 1850 and paper 3 links

parish-level data from the Tabular Commission to information in the 1900

population census. Additional city-level population data in papers 1 and 4 is

drawn from Lilja (1996) and Nilsson (1992), which report adjusted

population data based on poll-tax registers (mantalslängder) and the

population statistics maintained by the Tabular Commission.

Around the mid-19

_{century, the collection and reporting of official}

statistics broadened reflecting the increased complexity of the Swedish

economy. Statistics Sweden published a wide range of data in the series

Bidrag till Sveriges officiella statistik (BiSOS) that in its 23 different annual

volumes contains information on, for example, agricultural production,

banking activity, crime, domestic and foreign trade, education, industry, and

poor relief (http://www.scb.se/bisos). Paper 1 uses data from the agricultural

censuses to measure yields (Bidrag till Sveriges officiella statistik. N,

Jordbruk och boskapsskötsel), paper 2 uses data from the educational

statistics to track the expansion of elementary education (Bidrag till Sveriges

officiella statistik. P, Undervisningsväsendet) and information on educational

expenditure and the distribution of voting rights from the series on municipal

finances and poor relief to examine the link between the spread of the

15

_{This section draws on “Tabellverket - history and information about Tabellverket” that has}